This project demonstrates a clean, production-style dbt setup modeling a simple e-commerce system.

It follows best practices for modular, environment-agnostic dbt projects, including raw source management, staging, marts, tests, and documentation.

This tutorial walks through building a simple, production-quality modern data platform:

• Ingest raw data
• Transform it using dbt
• Analyze with business marts
• Visualize key metrics

You will need:

• Google Cloud account with BigQuery enabled (free tier is sufficient)
• dbt CLI installed locally (or optionally a dbt Cloud account)
• Tableau Public account (optional for visualization)

Recommended tools:

• Git
• VS Code (or another code editor)

The project is organized into three main layers:

• Raw sources: simulated customers and orders data (in dev_raw dataset)
• Staging models: light cleaning and standardization
• Mart models: business-ready metrics including Customer Lifetime Value and Monthly Revenue Trends

• Logical project structure: models/staging/, models/marts/
• Raw data sources declared using source()
• Staging models (stg_*) clean and standardized
• Business models (dim_*, fact_*) built with ref()
• Tests (unique, not_null) declared in schema.yml
• All models and columns documented
• Environment-agnostic: no hardcoded project/schema names
• Separate targets for seeds (seed_raw) and models (dev)
• Clean local runs: dbt run, dbt test, dbt docs generate
• Secure: No credentials or secrets in the repo

models/
├── staging/
│   ├── schema.yml
│   ├── stg_customers.sql
│   └── stg_orders.sql
├── marts/
│   ├── schema.yml
│   ├── dim_customers.sql
│   └── fact_orders.sql
seeds/
├── customers.csv
├── orders.csv
├── schema.yml
dbt_project.yml
README.md
deploy.sh

git clone https://github.com/tstrall/analytics.git
cd analytics

1. Create a new dataset in BigQuery named dev_raw.
2. Create another dataset named dbt_yourname (e.g., dbt_tstrall) for dbt models.
3. Enable BigQuery Storage API if prompted.

Create or edit ~/.dbt/profiles.yml with the following structure:

analytics:
  outputs:
    dev:
      type: bigquery
      method: service-account
      project: your-gcp-project-id
      dataset: dbt_yourname
      keyfile: path/to/your/service_account.json
      threads: 4

    seed_raw:
      type: bigquery
      method: service-account
      project: your-gcp-project-id
      dataset: dev_raw
      keyfile: path/to/your/service_account.json
      threads: 4

  target: dev

Replace:

• your-gcp-project-id with your GCP project ID
• dbt_yourname with your dataset for models
• path/to/your/service_account.json with your service account key path

Test your connection:

dbt debug

This project provides seed files containing dummy raw data. To load them into BigQuery:

dbt --target seed_raw seed

This loads tables into your dev_raw dataset.

dbt --target seed_raw test --select path:seeds/

dbt run

dbt test --select path:models/

dbt docs generate
dbt docs serve

Access documentation locally at http://localhost:8080.

You can deploy everything automatically:

./deploy.sh

This will:

• Seed raw data
• Test seeds
• Build models
• Test models
• Generate documentation

For a full step-by-step deployment guide, see DEPLOYMENT.md.

• View Live Dashboard on Tableau Public
• Tableau Files and Documentation

Note:
This Tableau dashboard reflects a snapshot of the dbt-transformed data as of the publish date.
Tableau Public dashboards do not maintain live BigQuery connections.

This project is built first for BigQuery. Future enhancements include:

• Snowflake version
• Automated ingestion with Fivetran
• Additional visualization dashboards

• dbt — Data transformation and modeling
• Google BigQuery — Cloud data warehouse
• Tableau — Data visualization
• GitHub — Source control and project management

Built by Ted Strall — strall.com

# Seed data into dev_raw
dbt --target seed_raw seed

# Test seeds
dbt --target seed_raw test --select path:seeds/

# Build models into dbt_tstrall
dbt run

# Test models
dbt test --select path:models/

# Generate docs
dbt docs generate

# Serve docs
 dbt docs serve

Version v1.0.0 — Initial dbt + BigQuery + Tableau stack

• dbt transformations and tests
• Automated deploy script
• Tableau dashboard for Top Customers and Revenue Trends